How tech is transforming the intelligence industry

At a conference on the future challenges of intelligence organizations held in 2018, former Director of National Intelligence Dan Coats argued that he transformation of the American intelligence community must be a revolution rather than an evolution. The community must be innovative and flexible, capable of rapidly adopting innovative technologies wherever they may arise.

Intelligence communities across the Western world are now at a crossroads: The growing proliferation of technologies, including artificial intelligence, Big Data, robotics, the Internet of Things, and blockchain, changes the rules of the game. The proliferation of these technologies – most of which are civilian, could create data breaches and lead to backdoor threats for intelligence agencies. Furthermore, since they are affordable and ubiquitous, they could be used for malicious purposes.

The technological breakthroughs of recent years have led intelligence organizations to challenge the accepted truths that have historically shaped their endeavors. The hierarchical, compartmentalized, industrial structure of these organizations is now changing, revolving primarily around the integration of new technologies with traditional intelligence work and the redefinition of the role of the humans in the intelligence process.

Take for example Open-Source Intelligence (OSINT) – a concept created by the intelligence community to describe information that is unclassified and accessible to the general public. Traditionally, this kind of information was inferior compared to classified information; and as a result, the investments in OSINT technologies were substantially lower compared to other types of technologies and sources. This is changing now; agencies are now realizing that OSINT is easy to acquire and more beneficial, compared to other – more challenging – types of information.

Yet, this understanding trickle down solely, as the use of OSINT by intelligence organizations still involves cumbersome processes, including slow and complex integration of unclassified and classified IT environments. It isn’t surprising therefore that intelligence executives – for example the Head of State Department’s Intelligence Arm or the nominee to become the Director of the National Reconnaissance Office – recently argued that one of the community’s grandest challenges is the quick and efficient integration of OSINT in its operations.

Indeed, technological innovations have always been central to the intelligence profession. But when it came to processing, analyzing, interpreting, and acting on intelligence, however, human ability – with all its limitations – has always been considered unquestionably superior. That the proliferation of data and data sources are necessitating a better system of prioritization and analysis, is not questionable. But who should have a supremacy? Humans or machines?

A man crosses the Central Intelligence Agency (CIA) seal in the lobby of CIA Headquarters in Langley, Virginia, on August 14, 2008. (Photo: SAUL LOEB/AFP/Getty Images)

Big data comes for the spy business

The discourse is tempestuous. Intelligence veterans claim that there is no substitute for human judgment. They argue that artificial intelligence will never be capable of comprehending the full spectrum of considerations in strategic decision-making, and that it cannot evaluate abstract issues in the interpretation of human behavior. Machines can collect data and perhaps identify patterns, but they will never succeed in interpreting reality as do humans. Others also warn of the ethical implications of relying on machines for life-or-death situations, such as a decision to go to war.

In contrast, techno-optimists claim that human superiority, which defined intelligence activities over the last century, is already bowing to technological superiority. While humans are still significant, their role is no longer exclusive, and perhaps not even the most important in the process. How can the average intelligence officer cope with the ceaseless volumes of information that the modern world produces?

From 1995 to 2016, the amount of reading required of an average US intelligence researcher, covering a low-priority country, grew from 20,000 to 200,000 words per day. And that is just the beginning. According to forecasts, the volume of digital data that humanity will produce in 2025 will be ten times greater than is produced today. Some argue this volume can only be processed – and even analyzed – by computers.

Of course, the most ardent advocates for integration of machines into intelligence work are not removing human involvement entirely; even the most skeptical do not doubt the need to integrate artificial intelligence into intelligence activities. The debate centers on the question of who will help whom: machines in aid of humans or humans in aid of machines.

Most insiders agree that the key to moving intelligence communities into the 21st century lies in breaking down inter- and intra-organizational walls, including between
the services within the national security establishment; between the public sector, the private sector, and academia; and between intelligence services of different countries.

It isn’t surprising therefore that the push toward technological innovation is a part of the current intelligence revolution. The national security establishment already recognizes that the private sector and academia are the main drivers of technological innovation.

Alexander Karp, chief executive officer and co-founder of Palantir Technologies Inc., walks the grounds after the morning sessions during the Allen & Co. Media and Technology Conference in Sun Valley, Idaho, U.S., on Thursday, July 7, 2016. Billionaires, chief executive officers, and leaders from the technology, media, and finance industries gather this week at the Idaho mountain resort conference hosted by investment banking firm Allen & Co. Photographer: David Paul Morris/Bloomberg via Getty Images

Private services and national intelligence

In the United States there is dynamic cooperation between these bodies and the security community, including venture capital funds jointly owned by the government and private companies.

Take In-Q-Tel – a venture capital fund established 20 years ago to identify and invest in companies that develop innovative technology which serves the national security of the United States, thus positioning the American intelligence community at the forefront of technological development. The fund is an independent corporation, which is not subordinate to any government agency, but it maintains constant coordination with the CIA, and the US government is the main investor.

It’s most successful endeavor, which has grown to become a multi-billion company though somewhat controversial, is Palantir, a data-integration and knowledge management provider. But there are copious other startups and more established companies, ranging from sophisticated chemical detection (e.g. 908devices), automated language translations (e.g. Lilt), and digital imagery (e.g. Immersive Wisdom) to sensor technology (e.g. Echodyne), predictive analytics (e.g. Tamr) and cyber security (e.g. Interset).

Actually, a significant part of intelligence work is already being done by such companies, small and big. Companies like Hexagon, Nice, Splunk, Cisco and NEC offer intelligence and law enforcement agencies a full suite of platforms and services, including various analytical solutions such as video analytics, identity analytics, and social media analytics . These platforms help agencies to obtain insights and make predictions from the collected and historic data, by using real-time data stream analytics and machine learning. A one-stop-intelligence-shop if you will.

Another example of government and non-government collaboration is the Intelligence Advanced Research Projects Activity (IARPA) – a nonprofit organization which reports to the Director of National Intelligence (DNI). Established in 2006, IARPA finances advanced research relevant to the American intelligence community, with a focus on cooperation between academic institutions and the private sector, in a broad range of technological and social sciences fields. With a relatively small annual operational budget of around $3bn, the fund gives priority to multi-year development projects that meet the concrete needs of the intelligence community. The majority of the studies supported by the fund are unclassified and open to public scrutiny, at least until the stage of implementation by intelligence agencies.

Image courtesy of Bryce Durbin/TechCrunch

Challenging government hegemony in the intelligence industry 

These are all exciting opportunities; however, the future holds several challenges for intelligence agencies:

First, intelligence communities lose their primacy over collecting, processing and disseminating data. Until recently, the organizations Raison D’etre was, first and foremost, to obtain information about the enemy, before said enemy could disguise that information.

Today, however, a lot of information is available, and a plethora of off-the-shelf tools (some of which are free) allow all parties, including individuals, to collect, process and analyze vast amounts of data. Just look at IBM’s i2 Analyst’s Notebook, which gives analysts, for just few thousand dollars, multidimensional visual analysis capabilities so they can quickly uncover hidden connections and patterns in data. Such capacities belonged, just until recently, only to governmental organizations.

A second challenge for intelligence organizations lies in the nature of the information itself and its many different formats, as well as in the collection and processing systems, which are usually separate and lacking standardization. As a result, it is difficult to merge all of the available information into a single product. For this reason, intelligence organizations are developing concepts and structures which emphasize cooperation and decentralization.

The private market offers a variety of tools for merging information; ranging from simple off-the-shelf solutions, to sophisticated tools that enable complex organizational processes. Some of the tools can be purchased and quickly implemented – for example, data and knowledge sharing and management platforms – while others are developed by the organizations themselves to meet their specific needs.

The third challenge relates to the change in the principle of intelligence prioritization. In the past, the collection of information about a given target required a specific decision to do so and dedicated resources to be allocated for that purpose, generally at the expense of allocation of resources to a different target. But in this era of infinite quantities of information, almost unlimited access to information, advanced data storage capabilities and the ability to manipulate data, intelligence organizations can now collect and store information on a massive scale, without the need to immediately process it – rather, it may be processed as required.

This development leads to other challenges, including: the need to pinpoint the relevant information when required; to process the information quickly; to identify patterns and draw conclusions from mountains of data; and to make the knowledge produced accessible to the consumer. It is therefore not surprising that most of the technological advancements in the intelligence field respond to these challenges, bringing together technologies such as big data with artificial intelligence, advanced information storage capabilities and advanced graphical presentation of information, usually in real time.

Lastly, intelligence organizations are built and operate according to concepts developed at the peak of the industrial era, which championed the principle of the assembly line, which are both linear and cyclical. The linear model of ​​the intelligence cycle – collection, processing, research, distribution and feedback from the consumer – has become less relevant. In this new era, the boundaries between the various intelligence functions and between the intelligence organizations and their eco-system are increasingly blurred.

 

The brave new world of intelligence

A new order of intelligence work is therefore required, and therefore intelligence organizations are currently in the midst of a redefinition process. Traditional divisions – e.g. between collection and research; internal security organizations and positive intelligence; and public and private sectors – all become obsolete. This is not another attempt to carry out structural reforms: there is a sense of epistemological rupture which requires a redefinition of the discipline, the relationships that intelligence organizations have with their environments – from decision makers to the general public – and the development of new structures and conceptions.

And of course, there are even wider concerns; legislators need to create a legal framework that accurately incorporates the assessments based on data in a way that takes the predictive aspects of these technologies into account and still protects the privacy and security rights of individual citizens in nation states that have a respect for those concepts.

Despite the recognition of the profound changes taking place around them, today’s intelligence institutions are still built and operate in the spirit of Cold War conceptions. In a sense, intelligence organizations have not internalized the complexity that characterizes the present time – a complexity which requires abandoning the dichotomous (within and outside) perception of the intelligence establishment, as well as the understanding of the intelligence enterprise and government bodies as having a monopoly on knowledge; concepts that have become obsolete in an age of decentralization, networking and increasing prosperity.

Although some doubt the ability of intelligence organizations to transform and adapt themselves to the challenges of the future, there is no doubt that they must do so in this era in which speed and relevance will determine who prevails.

Duo’s Wendy Nather to talk security at TC Sessions: Enterprise

When it comes to enterprise security, how do you move fast without breaking things?

Enter Duo’s Wendy Nather, who will join us at TC Sessions: Enterprise in San Francisco on September 5, where we will get the inside track on how to keep enterprise networks secure without slowing growth.

Nather is head of advisory CISOs at Duo Security, a Cisco company, and one of the most respected and trusted voices in the cybersecurity community as a regular speaker on a range of topics, from threat intelligence to risk analysis, incident response, data security and privacy issues.

Prior to her role at Duo, she was the research director at the Retail ISAC, and served as the research director of the Information Security Practice at independent analyst firm 451 Research.

She also led IT security for the EMEA region of the investment banking division of Swiss Bank Corporation — now UBS.

Nather also co-authored “The Cloud Security Rules,” and was listed as one of SC Magazine’s Women in IT Security “Power Players” in 2014.

We’re excited to have Nather discuss some of the challenges startups and enterprises face in security — threats from both inside and outside the firewall. Companies large and small face similar challenges, from keeping data in to keeping hackers out. How do companies navigate the litany of issues and threats without hampering growth?

Who else will we have onstage, you ask? Good question! We’ll be joined by some of the biggest names and the smartest and most prescient people in the industry, including Bill McDermott at SAP, Scott Farquhar at Atlassian, Julie Larson-Green at Qualtrics, Aaron Levie at Box and Andrew Ng at Landing AI and many, many more. See the whole agenda right here.

Early-bird tickets are on sale right now! For just $249 you can see Nather and these other awesome speakers live at TC Sessions: Enterprise. But hurry, early-bird sales end on August 9; after that, prices jump up by $100. Book here.

If you’re a student on a budget, don’t worry, we’ve got a super-reduced ticket for just $75 when you apply for a student ticket right here.

Enterprise-focused startups can bring the whole crew when you book a Startup Demo table for just $2,000. Each table gives you a primo location to be seen by attendees, investors and other sponsors, in addition to four tickets to enjoy the show. We only have a limited amount of demo tables and we will sell out. Book yours here.

Hyundai takes minority stake in self-driving car startup Aurora

Hyundai Motor Group has invested in Aurora, the latest sign that the scope of the year-old partnership between the automaker and self-driving car startup has expanded.

Aurora and Hyundai didn’t disclose terms of the investment. However, picking part new details of its Series B funding round and speaking to sources within the industry, Hyundai’s investment is below $30 million.

Aurora announced in February that it had raised more than $530 million in a Series B round that was led by Sequoia Capital and included “significant investment” from Amazon and T. Rowe Price Associates. Since then, that Series B round has expanded to more than $600 million with new investment from Hyundai, Baillie Gifford and the Canada Pension Plan Investment Board, TechCrunch has learned.

To date, Aurora has raised more than $700 million, a figure that includes its seed round Series A round of $90 million.

Hyundai’s stake in Aurora is an affirmation of the company and their working relationship. But it’s just one measure. What Aurora is actually doing matters as much.

When the partnership was first announced in January 2018, the details of the relationship were scant. New information reveals that Aurora has been working with Hyundai and Kia for the last year to integrate its “Driver” into Hyundai’s flagship fuel cell vehicle NEXO.

Aurora says it will expand research and development of a self-driving platform for a wide range of Hyundai and Kia’s models.

Aurora, which launched in January 2017, works with companies like Hyundai, Byton, and until more recently Volkswagen, to design and develop a package of sensors, software, and data services needed to deploy autonomous vehicles. The company describes this “full stack,” (an industry parlance) the Aurora Driver.

Aurora, like many of its competitors, are focused on Level 4 autonomous systems with an eye toward Level 5. Level 4 is a designation by SAE, the automotive engineering association, for autonomous vehicles that take over all driving in certain conditions. In Level 5 autonomy, the vehicle is self-driving in all situations.

About a year after Aurora’s official launch date, the company announced partnerships with Hyundai and Volkswagen, followed by a Series A funding raise that resulted in two new board members — LinkedIn co-founder and Greylock partner Reid Hoffman and Mike Volpi, former chief strategy officer at Cisco and general partner at Index Ventures.

Volkswagen has since ended its partnership with Aurora. Meanwhile, Fiat Chrysler Automobiles has announced a collaboration with Aurora to to develop self-driving commercial vehicles. The partnership with FCA will focus on integrating Aurora’s technology into the automaker’s line of Ram Truck commercial vehicles, a portfolio that includes cargo vans and trucks. The deal could extend to FCA’s Fiat Professional brand as well, TechCrunch has learned.

Newly public CrowdStrike wants to become the Salesforce of cybersecurity

Like many good ideas, CrowdStrike, a seller of subscription-based software that protects companies from breaches, began as a few notes scribbled on a napkin in a hotel lobby.

The idea was to leverage new technology to create an endpoint protection platform powered by artificial intelligence that would blow incumbent solutions out of the water. McAfee, Palo Alto Networks and Symantec, long-time leaders in the space, had been too slow to embrace new technologies and companies were suffering, the CrowdStrike founding team surmised.

Co-founders George Kurtz and Dmitri Alperovitch, a pair of former McAfee executives, weren’t strangers to legacy cybersecurity tools. McAfee had for years been a dominant player in endpoint protection and antivirus. At least, until the emergence of cloud computing.

Since 2012, CrowdStrike’s Falcon Endpoint Protection platform has been pushing those incumbents into a new era of endpoint protection. By helping enterprises across the globe battle increasingly complex attack scenarios more efficiently, CrowdStrike, as well as other fast-growing cybersecurity upstarts, has redefined company security standards much like Salesforce redefined how companies communicate with customers.

“I think we had the foresight that [CrowdStrike] was going to be a foundational element for security,” CrowdStrike chief executive officer George Kurtz told TechCrunch this morning. The full conversation can be read further below.

CrowdStrike co-founder and CEO George Kurtz.

Social Capital reincarnated

Nine months ago, the once high-flying venture capital fund Social Capital made the bold decision to stop accepting outside capital and operate as a family office, in essence.

The co-founder of the outfit, brazen billionaire and early Facebook executive Chamath Palihapitiya, pledged to upend his investment strategy and make fewer but much larger investments as a means to improve his returns. Naturally, a near-complete exodus of Social Capital’s venture capitalists followed.

Today, the firm’s three founders, Palihapitiya, Mamoon Hamid and Ted Maidenberg, have gone their separate ways. Palihapitiya is rewriting the Social Capital playbook, Hamid is busy reinvigorating Kleiner Perkins and Maidenberg is building on top of the data-driven strategy and proprietary software dubbed “Magic 8-Ball” he built at Social Capital, with a new firm called Tribe Capital.

Quietly, Tribe Capital’s co-founders, Maidenberg and former Social Capital partners Arjun Sethi and Jonathan Hsu, have deployed millions of dollars in Social Capital portfolio companies like Slack and Carta, hired several former Social Capital employees and flexed a data-first approach that looks pretty damn familiar.  

Data or bust

SAN FRANCISCO, CA – OCTOBER 19: Founder/CEO of Social Capital, Chamath Palihapitiya, speaks onstage during “The State of the Valley: Where’s the Juice?” (Photo by Michael Kovac/Getty Images for Vanity Fair)

Social Capital began laying the foundation for a data-driven approach to investing years ago. Now, Tribe Capital is doubling down.

From its founding in 2011, Social Capital established itself as a contrarian fund out to “fix capitalism.” Its strategy and reputation as an up-and-comer unafraid of new tricks earned it stakes in Slack, SurveyMonkey, Box, Bust and many other admirable upstarts.

As the firm matured, its partners experimented. In 2016, its early-stage investment team made the daring choice to rely on data rather than gut-feel alone to make its investment decisions, confronting a timeworn ideology that the best VCs have a special skill-set that enables them to spot future unicorns.

Using an operating system for early-stage investing dubbed “capital-as-a-service” and the growth and data analysis tool Magic 8-Ball — a sort of QuickBooks for startup data — Social Capital forwent the traditional pitch process and rapidly evaluated thousands of companies on the basis of metrics and achievements alone.

Palihapitiya, Maidenberg, Hamid and the other members of the partnership were on a mission to do venture the right way. Until they weren’t.

“I found us incrementally drifting away from our core mission, and our strategy was increasingly that of a traditional investment firm,” Palihapitiya wrote last year. “It became harder to take the risks we took in 2011 and it became easier to play the same game as every other VC.”

At its peak, Social Capital employed a team of 80. Once Palihapitiya confirmed his intent to transition the firm away from venture, the team began to shrink, fast. Today, the firm employs 30, including partners Ray Ko, Andy Artz and Jay Zaveri, as well as principal Alex Danco. One-third of that number were hired after the big pivot.

The Social Capital diaspora 

Social Capital co-founder Mamoon Hamid left the fund in 2017 for Kleiner Perkins.

Social Capital’s former investors have since identified their second acts.

In the last year, Sakya Duvvuru, a former partner, founded Nellore Capital Management, and Carl Anderson, another former partner, started Marcho Partners.

Tony Bates joined Genesys as its CEO, Mike Ghaffary accepted a general partner role at Canvas Ventures, Ashley Carroll is consulting full-time, Kristen Spohn says she is still exploring opportunities, Adam Nelson joined South Park Commons as a venture partner and Tejinder Gill joined Collaborative Fund as a principal.

Hamid, for his part, resolved to re-establish Kleiner Perkins’ once-stellar reputation.

“Kleiner Perkins was a firm that was in desperate need of a change of its own,” Hamid tells TechCrunch. “It was a unique opportunity and I was about to turn 40. I thought, there is one thing I wanted to do in my career that I hadn’t done before and that was to turn around one of the best venture firms of all time.”

Hamid’s August 2017 departure from Social Capital represented the beginning of the end of the partnership. Though Hamid, a co-founder and leading dealmaker, asserts turmoil at the firm began after his exit. 

Nine months after Hamid made the call to move on, Arjun Sethi, who once led Social Capital’s early-stage investment team, made the same call as did Maidenberg and Hsu. Simultaneously, growth equity chief Tony Bates and vice chairman Marc Mezvinsky were said to be departing.

The mass exodus continued, culminating in Palihapitiya’s final declaration: Social Capital was finished with venture capital.

‘Magic 8-Ball’ — reborn

Maidenberg, Sethi and Hsu built Tribe Capital in the image of Social Capital. With similar DNA, the three men are attempting to upgrade an early-stage investment strategy they not only created, but nearly perfected.

“Those guys did a very good job working for me,” Palihapitiya tells TechCrunch. “I’m super proud to see them launch their own venture fund. It was a really important, defining experience for me; I hope they have the same level of success, if not more.”

But where Social Capital was mission-driven, regularly backing healthcare and education businesses, Tribe Capital makes no such claim. And where Social Capital leaned on data to inform its investment thesis, Tribe is putting its full weight into it.

We are believers that it’s hard to do a lot of things well, so we wanted to focus on one thing we are good at: early-stage venture with the approach of recognizing early-stage product-market fit,” Hsu tells TechCrunch. “At Social Capital we did that, but we did 30 other things, too.”

In total, seven former Social Capital investors and employees are working on Tribe. Georgia Kinne, a former Social Capital executive assistant, leads operations. Two former Social Capital data scientists, Jake Ellowitz and Brendan Moore, joined Tribe in the same role. And Alexander Chee, Social Capital’s former head of product development, is on board as an entrepreneur-in-residence.

Tribe won’t say how much capital they have raised yet or how exactly their three funds are structured, aside from confirming that only one is operating as a traditional venture fund. Paperwork filed with the U.S. Securities and Exchange Commission in late April, however, confirms a $150 million target for the debut venture effort. 

It’s been a year since Tribe began investing. In that time, it’s put money in Slack, Front, Cover and SFox. Most recently, it participated in Carta’s $300 million Series E, which valued the business at $1.7 billion. All of these companies were previously backed by Social Capital.

Tribe is making deals of all shapes and sizes across industries, with a particular focus on enterprise, fintech and SaaS startups. In addition to deploying heftier sums to late-stage businesses like Slack, Tribe has made 10 seed bets of roughly $25,000 each, leveraging its data platform to make investment calls.

“The income statement and balance sheet are the lingua franca for an established company to communicate the financial health of its business,” Hsu writes. “These accounting concepts are often unhelpful when inspecting an unprofitable early-stage company. For a startup, what’s needed is a common quantitative language for what matters, namely, a quantitative framework for assessing product-market fit.”

Tribe’s quantitative framework is called Magic 8-Ball, a diligence tool for potential investments created by Maidenberg and Hsu during their Social Capital tenure. The tool measures product-market fit, growth trajectory and more of early-stage businesses, where, as Hsu mentions, financial data may be lacking.

“We use data like accountants; it’s not a magical AI machine,” Hsu said. “If other firms want to copy, by all means, they can try. We aren’t here to be antagonistic, we are here to be partners to founders and other investors.”

So far, Magic 8-Ball has poured through data provided by some 200 companies, with plans to hit 1,000 per year. In total, Tribe has deployed $100 million.

Tribe’s 8-Ball tool is said to be much more complex than the earlier model, according to a source with knowledge of the platform. It’s like when Yahoo engineers Jan Koum and Brian Acton left the search and email giant to build something even better, the source, who asked not to be named, said. That business became the messaging powerhouse WhatsApp.

Hamid, who’s not affiliated with Tribe but aware of their investment strategy, made a similar comparison.

“It’s like if you’re an engineer at Cisco working on WebEx,” Hamid tells TechCrunch. “You’re a great engineer but you can do better, you can [do your own] company. Guess what? That’s Zoom. That’s Eric Yuan . And Zoom is worth $20 billion and WebEx was worth $3 billion. That’s pretty. That’s the story of Silicon Valley. That’s creative disruption.”

Hamid, however, was careful to point out the differences between Social Capital and Tribe. The DNA may be similar but they aren’t identical.

Social Capital represented a new kind of venture firm in favor of creative disruption. Tribe Capital represents a second go, a sort of Social Capital 2.0 sans Chamath Palihapitiya.

Bogged down by the conflict surrounding its leader’s flair for controversy, Social Capital wasn’t set up to succeed. The Magic-8 Ball, on the other hand, may be just right.

“Why did we get back together instead of going elsewhere? That is a reasonable question,” Hsu said. “We had good job offers but we had a viewpoint of the world that we wanted to keep working on together.”

Diving deep into Africa’s blossoming tech scene

Jumia may be the first startup you’ve heard of from Africa. But the e-commerce venture that recently listed on the NYSE is definitely not the first or last word in African tech.

The continent has an expansive digital innovation scene, the components of which are intersecting rapidly across Africa’s 54 countries and 1.2 billion people.

When measured by monetary values, Africa’s tech ecosystem is tiny by Shenzen or Silicon Valley standards.

But when you look at volumes and year over year expansion in VC, startup formation, and tech hubs, it’s one of the fastest growing tech markets in the world. In 2017, the continent also saw the largest global increase in internet users—20 percent.

If you’re a VC or founder in London, Bangalore, or San Francisco, you’ll likely interact with some part of Africa’s tech landscape for the first time—or more—in the near future.

That’s why TechCrunch put together this Extra-Crunch deep-dive on Africa’s technology sector.

Tech Hubs

A foundation for African tech is the continent’s 442 active hubs, accelerators, and incubators (as tallied by GSMA). These spaces have become focal points for startup formation, digital skills building, events, and IT activity on the continent.

Prominent tech hubs in Africa include CcHub in Nigeria, Pan-African incubator MEST, and Kenya’s iHub, with over 200 resident members. More of these organizations are receiving funds from DFIs, such as the World Bank, and aid agencies, including France’s $76 million African tech fund.

Blue-chip companies such as Google and Microsoft are also providing money and support. In 2018 Facebook opened its own Hub_NG in Lagos with partner CcHub, to foster startups using AI and machine learning.

Rumpus, the collaborative toolkit from Oblong Industries, is now available on Webex

In a previous life, John Underkoffler spent his days in Los Angeles dreaming up all of the possible ways men and machines would interact as a science adviser on films like Minority Report.

Now, he designs those systems for the real world through his company Oblong Industries, which has labored to create a full stack of collaborative tools for business users that are every bit as high-tech as the one’s Underkoffler dreamt for the silver screen.

The first bolt in the quiver of tools that Underkoffler began building out over the course of 15 years spent at MIT’s Media Lab was Mezzanine. A multipurpose collaborative platform that allowed business users to share documents and interact in real time through a powerful combination of videoconferencing hardware and software.

In the age of Zoom though, Oblong’s tools have become more lightweight, and the company is steadily adding multi-share capabilities to platforms other than its own. That new gaggle of collaboration tools launched under the moniker of Rumpus, and Oblong has been partnering with different video services to add its services to their own.

The latest to get the Rumpus treatment is Cisco Webex.  Now Cisco’s videoconferencing customers will get access to Rumpus’ personal cursors that point and emphasize content on shared screens, presence indicators to show who is looking where and at what, and emoji reactions to provide feedback without disrupting the flow of a meeting.

The company’s tools enable all of the users in a meeting to share their screens without competing for screen time.

“We’ve worked closely with Cisco over the last year to bring the capabilities of our flagship product, Mezzanine, to the Cisco suite of enterprise solutions for meetings paces. So as we completed Oblong’s own set of content-first collaboration offerings by building out Rumpus for pure-virtual work, it was obvious that Webex should be among the first conferencing solutions to be directly integrated,” said Underkoffler in a statement. “We’re thrilled to bring . the next level of engagement and productivity to millions of Webex users when their meetings require more than basic video and messaging.”

Rumpus is currently available for free to Mac computer users with Windows support coming soon.

Fastly pops in public offering showing that there’s still money for tech IPOs

Shares of Fastly, the service that’s used by websites to ensure that they can load faster, have popped in its first hours of trading on the New York Stock Exchange.

The company, which priced its public offering at around $16 — the top of the estimated range for its public offering — have risen more than 50% since their debut on public markets to trade at $25.01.

It’s a sharp contrast to the public offering last week from Uber, which is only just now scratching back to its initial offering price after a week of trading underwater, and an indicator that there’s still some open space in the IPO window for companies to raise money on public markets, despite ongoing uncertainties stemming from the trade war with China.

Compared with other recent public offerings, Fastly’s balance sheet looks pretty okay. Its losses are narrowing (both on an absolute and per-share basis according to its public filing), but the company is paying more for its revenue.

San Francisco-based Fastly competes with companies that include Akamai, Amazon, Cisco and Verizon, providing data centers and a content-distribution service to deliver videos from companies like The New York Times, Ticketmaster, New Relic and Spotify.

Last year, the company reported revenues of $144.6 million and a net loss of $30.9 million, up from $104.9 million in revenue and $32.5 million in losses in the year ago period. Revenue was up more than 38% and losses narrowed by 5% over the course of the year.

The outcome is a nice win for Fastly investors, including August Capital, Iconiq Strategic Partners, O’Reilly AlphaTech Ventures and Amplify Partners, which backed the company with $219 million in funding over the eight years since Artur Bergman founded the business in 2011.

Cisco open sources MindMeld conversational AI platform

Cisco announced today that it was open sourcing the MindMeld conversation AI platform, making it available to anyone who wants to use it under the Apache 2.0 license.

MindMeld is the conversational AI company that Cisco bought in 2017. The company put the technology to use in Cisco Spark Assistant later that year to help bring voice commands to meeting hardware, which was just beginning to emerge at the time.

Today, there is a concerted effort to bring voice to enterprise use cases, and Cisco is offering the means for developers to do that with the MindMeld tool set. “Today, Cisco is taking a big step towards empowering developers with more comprehensive and practical tools for building conversational applications by open-sourcing the MindMeld Conversational AI Platform,” Cisco’s head of machine learning Karthik Raghunathanw wrote in a blog post.

The company also wants to make it easier for developers to get going with the platform, so it is releasing the Conversational AI Playbook, a step-by-step guide book to help developers get started with conversation-driven applications. Cisco says this is about empowering developers, and that’s probably a big part of the reason.

But it would also be in Cisco’s best interest to have developers outside of Cisco working with and on this set of tools. By open sourcing them, the hope is that a community of developers, whether Cisco customers or others, will begin using, testing and improving the tools; helping it to develop the platform faster and more broadly than it could, even inside an organization as large as Cisco.

Of course, just because they offer it doesn’t necessarily automatically mean the community of interested developers will emerge, but given the growing popularity of voice-enabled used cases, chances are some will give it a look. It will be up to Cisco to keep them engaged.

Cisco is making all of this available on its own DevNet platform starting today.

Unshackled Ventures has $20M to invest exclusively in immigrant founders

Unshackled Ventures isn’t like other venture capital funds.

The firm invests in immigrant founders and helps them secure visas so they can ditch their corporate job and launch the startup of their dreams. Today, Unshackled is announcing its sophomore fund of $20 million, topping its debut effort by $15.5 million.

“The point is to take the burden off of founders because they are not immigration experts, they are experts at building satellites or extracting protein from plants,” Unshackled founding partner Nitin Pachisia told TechCrunch. “These are people that if you go to a workspace, you’ll see them show up on nights and weekends because they want to build something but they can’t.”

Immigrants looking to start their own businesses face a huge barrier. Take Jyoti Bansal for example. He famously waited seven years before launching AppDynamics, a business that later sold to Cisco for $3.7 billion days before its initial public offering. Why? Because as an Indian immigrant with H-1B visa status, he could work for startups but wasn’t legally allowed to start his own. It wasn’t until receiving an employment authorization document (EAD), a part of the green card process, that Bansal could finally found AppDynamics. If Bansal had the opportunity to pitch to Unshackled, which provides bespoke immigration solutions to each founder, he could have launched AppDynamics years prior.

Immigrant founders, according to a 2018 study by the National Foundation for American Policy, are responsible for 55 percent of U.S. billion-dollar companies, or “unicorns,” as they are known. Uber, SpaceX, WeWork, Palantir Technologies, Stripe, Slack, Moderna Therapeutics, Robinhood, Instacart, Houzz, Credit Karma, Tanium, Zoox and CrowdStrike all count at least one immigrant co-founder.

“The difference between success and failures is oftentimes who you know and when,” Unshackled founding partner Manan Mehta told TechCrunch. “We can bring those resources at just 1/200th the size of Andreessen Horowitz to immigrants at day zero.”

“We’re creating the best place for immigrants to start their companies,” he added. “And guess what? We’re keeping American innovation in America.”

Unshackled Ventures portfolio company Lily AI.

The firm was founded by Mehta, the son of immigrants, and Pachisia, an Indian immigrant, in 2015. Since then, the duo have written pre-seed checks to 31 companies with a 100 percent success rate in procuring visas to keep talent working in the U.S. Startups in its portfolio include the very recent Y Combinator graduate Career Karma, Starsky Robotics, Plutoshift, Togg, Hype, Lily AI and more.

“I didn’t think it was possible to start a company on a visa in the U.S., let alone scale one to hit the next major milestone so quickly,” Plutoshift founder Prateek Joshi said in a statement. “That all changed when we met the Unshackled team.”

Mehta and Pachisia say its startups have gone on to raise $54 million in follow-on investments from top investors like First Round Capital, NEA and Shasta.

In addition to supporting companies based in Silicon Valley, the investors search far and wide for aspiring immigrant founders, as well as respond to every single cold email they receive. Recently, they joined the Rise of the Rest tour, a trip hosted by Steve Case and JD Vance that showcases startups in underrepresented geographies, and they make frequent visits to college campuses across the U.S.

Unshackled’s limited partners include Bloomberg Beta, Jerry Yang’s AME Cloud Ventures and Emerson Collective.

“I think the name represents the feeling that you’re a little bit shackled to a framework or a policy that doesn’t necessarily encourage entrepreneurship,” Mehta said. “When if you take a step back, immigrants are probably more entrepreneurial than native-born people.”