Personal blog of Yzmir Ramirez

Electric-vehicle chargers, heads-up displays for soldiers and the Costco of weed were some of our favorites from prestigious startup accelerator Y Combinator’s Winter 2019 Demo Day 1. If you want to take the pulse of Silicon Valley, YC is the place to be. But with more than 200 startups presenting across two stages and two days, it’s tough to keep track.

You can check out our write-ups of all 85 startups that launched on Demo Day 1, and come back later for our full index and picks from Day 2. But now, based on feedback from top investors and TechCrunch’s team, here’s our selection of the top 10 companies from the first half of this Y Combinator batch, and why we picked each.

Ravn

Looking around corners is one of the most dangerous parts of war for infantry. Ravn builds heads-up displays that let soldiers and law enforcement see around corners thanks to cameras on their gun, drones or elsewhere. The ability to see the enemy while still being behind cover saves lives, and Ravn already has $490,000 in Navy and Air Force contracts. With a CEO who was a Navy Seal who went on to study computer science, plus experts in augmented reality and selling hardware to the Department of Defense, Ravn could deliver the inevitable future of soldier heads-up displays.

Why we picked Ravn: The AR battlefield is inevitable, but right now Microsoft’s HoloLens team is focused on providing mid-fight information, like how many bullets a soldier has in their clip and where their squad mates are. Ravn’s tech was built by a guy who watched the tragic consequences of getting into those shootouts. He wants to help soldiers avoid or win these battles before they get dangerous, and his team includes an expert in selling hardened tech to the U.S. government.

Middesk

It’s difficult to know if a business’ partners have paid their taxes, filed for bankruptcy or are involved in lawsuits. That leads businesses to write off $120 billion a year in uncollectable bad debt. Middesk does due diligence to sort out good businesses from the bad to provide assurance for B2B deals, loans, investments, acquisitions and more. By giving clients the confidence that they’ll be paid, Middesk could insert itself into a wide array of transactions.

Why we picked Middesk: It’s building the trust layer for the business world that could weave its way into practically every deal. More data means making fewer stupid decisions, and Middesk could put an end to putting faith in questionable partners.

Convictional

Convictional helps direct-to-consumer companies approach larger retailers more simply. It takes a lot of time for a supplier to build a relationship with a retailer and start selling their products. Convictional wants to speed things up by building a B2B self-service commerce platform that allows retailers to easily approach brands and make orders.

Why we picked Convictional: There’s been an explosion of D2C businesses selling everything from suitcases to shaving kits. But to drive exposure and scale, they need retail partners who’re eager not to be cut out of this growing commerce segment. Playing middleman could put Convictional in a lucrative position, while also making it a nexus of valuable shopping data.

Dyneti Technologies

Dyneti has invented a credit card scanner SDK that uses a smartphone’s camera to help prevent fraud by more than 50 percent and improve conversion for businesses by 5 percent. The business was started by a pair of former Uber employees, including CEO Julia Zheng, who launched the fraud analytics teams for Account Security and UberEATS. Dyneti’s service is powered by deep learning and works on any card format. In the two months since it launched, the company has signed contracts with Rappi, Gametime and others.

Why we picked Dyneti: Cybersecurity threats are growing and evolving, yet underequipped businesses are eager to do more business online. Dyneti is one of those fundamental B2B businesses that feels like Stripe — capable of bringing simplicity and trust to a complex problem so companies can focus on their product.

AmpUp

The “Airbnb for electric-vehicle chargers,” ampUp is preparing for a world in which the majority of us drive EVs — it operates a mobile app that connects a network of thousands of EV chargers and drivers. Using the app, an electric-vehicle owner can quickly identify an available and compatible charger, and EV charger owners can earn cash sharing their charger at their own price and their own schedule. The service is currently live in the Bay Area.

Why we picked ampUp: Electric vehicles are inevitable, but reliable charging is one of the leading fears dissuading people from buying. Rather than build out some massive owned network of chargers that will never match the distributed gas station network, ampUp could put an EV charger anywhere there’s someone looking to make a few bucks.

Flockjay

Flockjay operates an online sales academy that teaches job seekers from underrepresented backgrounds the skills and training they need to pursue a career in tech sales. The 12-week bootcamp offers trainees coaching and mentorship. The company has launched its debut cohort with 17 students, 100 percent of whom are already in job interviews and 40 percent of whom have already secured new careers in the tech industry.

Why we picked Flockjay: Unlike coding bootcamps that can require intense prerequisites, killer salespeople can be molded from anyone with hustle. Those from underrepresented backgrounds already know how to expertly sell themselves to attain opportunities others take for granted. Flockjay could provide economic mobility at a crucial juncture when job security is shaky.

Deel

Twenty million international contractors work with U.S. companies, but it’s difficult to onboard and train them. Deel handles the contracts, payments and taxes in one interface to eliminate paperwork and wasted time. Deel charges businesses $10 per contractor per month and a 1 percent fee on payouts, which earns it an average of $560 per contractor per year.

Why we picked Deel: The destigmatization of remote work is opening new recruiting opportunities abroad for U.S. businesses. But unless teams can properly integrate these distant staffers, the cost savings of hiring overseas are negated. As the globalization megatrend continues, businesses will need better HR tools.

Glide

There has been a pretty major trend toward services that make it easier to build web pages or mobile apps. Glide lets customers easily create well-designed mobile apps from Google Sheets pages. This not only makes it easy to build the pages, but simplifies the skills needed to keep information updated on the site.

Why we picked Glide: While desktop website makers is a brutally competitive market, it’s still not easy to make a mobile site if you’re not a coder. Rather than starting from a visual layout tool with which many people would still be unfamiliar, Glide starts with a spreadsheet that almost everyone has used. And as the web begins to feel less personal with all the brands and influencers, Glide could help people make bespoke apps that put intimacy and personality first.

Docucharm

The platform, co-founded by former Uber product manager Minh Tri Pham, turns documents into structured data a computer can understand to accurately automate document processing workflows and take away the need for human data entry. Docucharm’s API can understand various forms of documents (like paystubs, for example) and will extract the necessary information without error. Its customers include tax prep company Tributi and lending business Aspire.

Why we picked Docucharm: Paying high-priced, high-skilled workers to do data entry is a huge waste. And optical character recognition like Docucharm’s will unlock new types of businesses based on data extraction. This startup could be the AI layer underneath it all.

Flower Co

Flower Co provides memberships for cheaper weed sales and delivery. Most dispensaries cater to high-end customers and newbies that want expensive products and tons of hand-holding. In contrast, Flower Co caters to long-time marijuana enthusiasts who want huge quantities at low prices. They’re currently selling $200.000 in marijuana per month to 700 members. They charge $100 a year for membership, and take 10 percent on product sales.

Why we picked Flower Co: Marijuana is the next gold rush, a once-in-a-generation land-grab opportunity. Yet most marijuana merchants have focused on hyper-discerning high-end customers despite the long-standing popularity of smoking big blunts of cheap weed with a bunch of friends. For those who want to make cannabis consumption a lifestyle, and there will be plenty, Flower Co could become their wholesaler.

Honorable Mentions

Atomic Alchemy – Filling the shortage of nuclear medicine

Yourchoice – Omni-gender non-hormonal birth control

Prometheus – Turning CO2 into gas

Lumos – Medical search engine for doctors

Heart Aerospace – Regional electric planes

Boundary Layer Technologies – Super-fast container ships

Additional reporting by Kate Clark, Greg Kumparak and Lucas Matney

For months, the drama has been steady in the Pentagon’s decade-long, $10 billion JEDI cloud contract procurement process. This week the plot thickened when the DOD reported that it has found new evidence of a possible conflict of interest, and has reopened its internal investigation into the matter.

“DOD can confirm that new information not previously provided to DOD has emerged related to potential conflicts of interest. As a result of this new information, DOD is continuing to investigate these potential conflicts,” Elissa Smith, Department of Defense spokesperson told TechCrunch.

It’s not clear what this new information is about, but The Wall Street Journal reported this week that senior federal judge Eric Bruggink of the U.S. Court of Federal Claims ordered that the lawsuit filed by Oracle in December would be put on hold to allow the DOD to investigate further.

From the start of the DOD RFP process, there have been complaints that the process itself was designed to favor Amazon, and that were possible conflicts of interest on the part of DOD personnel. The DOD’s position throughout has been that it is an open process and that an investigation found no bearing for the conflict charges. Something forced the department to rethink that position this week.

Oracle in particular has been a vocal critic of the process. Even before the RFP was officially opened, it was claiming that the process unfairly favored Amazon. In the court case, it made the conflict part clearer, claiming that an ex-Amazon employee named Deap Ubhi had influence over the process, a charge that Amazon denied when it joined the case to defend itself. Four weeks ago something changed when a single line in a court filing suggested that Ubhi’s involvement may have been more problematic than the DOD previously believed.

At the time, I wrote:

In the document, filed with the court on Wednesday, the government’s legal representatives sought to outline its legal arguments in the case. The line that attracted so much attention stated, “Now that Amazon has submitted a proposal, the contracting officer is considering whether Amazon’s re-hiring Mr. Ubhi creates an OCI that cannot be avoided, mitigated, or neutralized.” OCI stands for Organizational Conflict of Interest in DoD lingo.

And Pentagon spokesperson Heather Babb told TechCrunch:

During his employment with DDS, Mr. Deap Ubhi recused himself from work related to the JEDI contract. DOD has investigated this issue, and we have determined that Mr. Ubhi complied with all necessary laws and regulations.

Whether the new evidence that DOD has found is referring to Ubhi’s rehiring by Amazon or not is not clear at the moment, but it has clearly found new evidence it wants to explore in this case, and that has been enough to put the Oracle lawsuit on hold.

Oracle’s court case is the latest in a series of actions designed to protest the entire JEDI procurement process. The Washington Post reported last spring that co-CEO Safra Catz complained directly to the president. The company later filed a formal complaint with the Government Accountability Office (GAO), which it lost in November when the department’s investigation found no evidence of conflict. It finally made a federal case out of it when it filed suit in federal court in December, accusing the government of an unfair procurement process and a conflict on the part of Ubhi.

The cloud deal itself is what is at the root of this spectacle. It’s a 10-year contract worth up to $10 billion to handle the DOD’s cloud business — and it’s a winner-take-all proposition. There are three out clauses, which means it might never reach that number of years or dollars, but it is lucrative enough, and could possibly provide inroads for other government contracts, that every cloud company wants to win this.

The RFP process closed in October and the final decision on vendor selection is supposed to happen in April. It is unclear whether this latest development will delay that decision.

By now you’ve probably heard of the Defense Department’s massive winner-take-all $10 billion cloud contract dubbed the Joint Enterprise Defense Infrastructure (or JEDI for short).
Star Wars references aside, this contract is huge, even by government standards.The Pentagon would like a single cloud vendor to build out its enterprise cloud, believing rightly or wrongly that this is the best approach to maintain focus and control of their cloud strategy.

Department of Defense (DOD) spokesperson Heather Babb tells TechCrunch the department sees a lot of upside by going this route. “Single award is advantageous because, among other things, it improves security, improves data accessibility and simplifies the Department’s ability to adopt and use cloud services,” she said.

Whatever company they choose to fill this contract, this is about modernizing their computing infrastructure and their combat forces for a world of IoT, artificial intelligence and big data analysis, while consolidating some of their older infrastructure. “The DOD Cloud Initiative is part of a much larger effort to modernize the Department’s information technology enterprise. The foundation of this effort is rationalizing the number of networks, data centers and clouds that currently exist in the Department,” Babb said.

Setting the stage

It’s possible that whoever wins this DOD contract could have a leg up on other similar projects in the government. After all it’s not easy to pass muster around security and reliability with the military and if one company can prove that they are capable in this regard, they could be set up well beyond this one deal.

As Babb explains it though, it’s really about figuring out the cloud long-term. “JEDI Cloud is a pathfinder effort to help DOD learn how to put in place an enterprise cloud solution and a critical first step that enables data-driven decision making and allows DOD to take full advantage of applications and data resources,” she said.

The single vendor component, however, could explain why the various cloud vendors who are bidding, have lost their minds a bit over it — everyone except Amazon, that is, which has been mostly silent, happy apparently to let the process play out.

The belief amongst the various other players, is that Amazon is in the driver’s seat for this bid, possibly because they delivered a $600 million cloud contract for the government in 2013, standing up a private cloud for the CIA. It was a big deal back in the day on a couple of levels. First of all, it was the first large-scale example of an intelligence agency using a public cloud provider. And of course the amount of money was pretty impressive for the time, not $10 billion impressive, but a nice contract.

For what it’s worth, Babb dismisses such talk, saying that the process is open and no vendor has an advantage. “The JEDI Cloud final RFP reflects the unique and critical needs of DOD, employing the best practices of competitive pricing and security. No vendors have been pre-selected,” she said.

Complaining loudly

As the Pentagon moves toward selecting its primary cloud vendor for the next decade, Oracle in particular has been complaining to anyone who will listen that Amazon has an unfair advantage in the deal, going so far as to file a formal complaint last month, even before bids were in and long before the Pentagon made its choice.

Somewhat ironically, given their own past business model, Oracle complained among other things that the deal would lock the department into a single platform over the long term. They also questioned whether the bidding process adhered to procurement regulations for this kind of deal, according to a report in the Washington Post. In April, Bloomberg reported that co-CEO Safra Catz complained directly to the president that the deal was tailor made for Amazon.

Microsoft hasn’t been happy about the one-vendor idea either, pointing out that by limiting itself to a single vendor, the Pentagon could be missing out on innovation from the other companies in the back and forth world of the cloud market, especially when we’re talking about a contract that stretches out for so long.

As Microsoft’s Leigh Madden told TechCrunch in April, the company is prepared to compete, but doesn’t necessarily see a single vendor approach as the best way to go. “If the DOD goes with a single award path, we are in it to win, but having said that, it’s counter to what we are seeing across the globe where 80 percent of customers are adopting a multi-cloud solution,” he said at the time.

He has a valid point, but the Pentagon seems hell bent on going forward with the single vendor idea, even though the cloud offers much greater interoperability than proprietary stacks of the 1990s (for which Oracle and Microsoft were prime examples at the time).

Microsoft has its own large DOD contract in place for almost a billion dollars, although this deal from 2016 was for Windows 10 and related hardware for DOD employees, rather than a pure cloud contract like Amazon has with the CIA.

It also recently released Azure Stack for government, a product that lets government customers install a private version of Azure with all the same tools and technologies you find in the public version, and could prove attractive as part of its JEDI bid.

Cloud market dynamics

It’s also possible that the fact that Amazon controls the largest chunk of the cloud infrastructure market, might play here at some level. While Microsoft has been coming fast, it’s still about a third of Amazon in terms of market size, as Synergy Research’s Q42017 data clearly shows.

The market hasn’t shifted dramatically since this data came out. While market share alone wouldn’t be a deciding factor, Amazon came to market first and it is much bigger in terms of market than the next four combined, according to Synergy. That could explain why the other players are lobbying so hard and seeing Amazon as the biggest threat here, because it’s probably the biggest threat in almost every deal where they come up against each other, due to its sheer size.

Consider also that Oracle, which seems to be complaining the loudest, was rather late to the cloud after years of dismissing it. They could see JEDI as a chance to establish a foothold in government that they could use to build out their cloud business in the private sector too.

10 years might not be 10 years

It’s worth pointing out that the actual deal has the complexity and opt-out clauses of a sports contract with just an initial two-year deal guaranteed. A couple of three-year options follow, with a final two-year option closing things out. The idea being, that if this turns out to be a bad idea, the Pentagon has various points where they can back out.

In spite of the winner-take-all approach of JEDI, Babb indicated that the agency will continue to work with multiple cloud vendors no matter what happens. “DOD has and will continue to operate multiple clouds and the JEDI Cloud will be a key component of the department’s overall cloud strategy. The scale of our missions will require DOD to have multiple clouds from multiple vendors,” she said.

The DOD accepted final bids in August, then extended the deadline for Requests for Proposal to October 9th. Unless the deadline gets extended again, we’re probably going to finally hear who the lucky company is sometime in the coming weeks, and chances are there is going to be lot of whining and continued maneuvering from the losers when that happens.

When the Department of Defense announced in March a 10-year winner-take-all cloud contract that could be worth up to $10 billion, it raised a few eyebrows. Last week, they clarified some of the conditions, and it turns out that much like a modern baseball free agent contract, there are a couple of points where the DoD can opt out of the deal.

In a press conference last week, chief Pentagon spokesperson Dana W. White indicated that the original contract award is for just two years. After that there are two additional options for five years and three years. The department can opt out after the first two years if it’s not working out, or seven years if it accepts the second option. Obviously if it takes the final option, that would add up to a full 10-year commitment.

Leigh Madden, who heads up Microsoft’s defense effort, says he believes Microsoft can win such a contract, but it isn’t necessarily the best approach for the DoD. “If the DoD goes with a single award path, we are in it to win, but having said that, it’s counter to what we are seeing across the globe where 80 percent of customers are adopting a multi-cloud solution,” Madden told TechCrunch.

White indicated that 46 companies have responded to the request for proposal, but it seems clear there are only a handful of companies that could handle a project of this scope. For starters, we have Amazon and Microsoft, along with Google, IBM and Oracle.

That said, White indicated that companies can band together and form a partnership, which means you could see some extremely strange bedfellows trying to form the equivalent of a rock supergroup with multiple players coming together to win the deal.

This development certainly opens up some interesting options for the vendors involved and creates a level of competition and alliances the likes of which the tech industry might have never seen. Whoever gets the contract, they get two years to prove they can do this, then they will be evaluated before getting a shot at the second five-year window.